Time is melting away for one of Antarctica (opens in new tab)'s biggest glaciers, and its rapid deterioration could end with the ice shelf's complete collapse in just a few years, researchers warned at a virtual press briefing on Monday (Dec. 13) at the annual meeting of the American Geophysical Union (AGU).
Super Collapse 3 Crack 45
Download Zip: https://7curioqconsme.blogspot.com/?jj=2vGLto
"And it could lead to even more sea-level rise, up to 10 feet [3 m], if it draws the surrounding glaciers with it," Scambos said in a statement (opens in new tab), referring to the weakening effect one ice shelf collapse can have on other nearby glaciers.
The effect would be somewhat like that of a car window "where you have a few cracks that are slowly propagating, and then suddenly you go over a bump in your car and the whole thing just starts to shatter in every direction," she said.
Some of the changes in Thwaites' ice are so swift and dramatic that scientists are watching them happen in real time, such as the appearance two years ago of a giant rift on the eastern ice shelf, Pettit said. A series of recent satellite images showed the lengthening crack heading right for the spot where the researchers had planned to set up their field site for the season. While the crack wasn't moving fast enough to threaten their field work that year, seeing its implacable advance was still a sobering moment; the researchers nicknamed the crack "the dagger," Pettit said at the briefing.
While the immediate prognosis is grim for Thwaites' ice shelf, the longterm forecast for the rest of the glacier is less certain. Should the shelf collapse, the glacier's flow will likely accelerate in its rush toward the ocean, with parts of it potentially tripling in speed; other chain reactions could also play a part in driving accelerated ice fracturing and melt, Scambos said at AGU. But the timeframe for those changes will be decades rather than a handful of years, according to the briefing.
The release of snow slab avalanches results from a succession of mechanical processes1,2. A failure is initiated in a highly porous weak snow layer buried beneath a cohesive snow slab, leading to mixed-mode and quasi-brittle crack propagation along the slope3. If the slope angle is larger than the weak layer friction angle, the slab eventually slides and releases4.
A simulation is presented of critical (green squares) and super-critical (red circles) crack lengths normalized by the elastic length Λ as a function of the normalized slope angle \((\psi-\phi^*)/\phi\) (ϕ, friction angle; \(\phi^*\), effective friction angle). Solid and dashed lines correspond to theoretical curves for effective friction angles of \(\phi^*=\phi\) and \(\phi^*=0.4\), respectively. The inset shows \(\phi^*/\phi\) as a function of the collapse amplitude h normalized by weak-layer thickness Dwl (data points correspond to an average of several simulations and error bars represent the s.d.).
Numerous numerical simulations were performed for different slope angles and mechanical properties of both the slab and weak layer (Methods) to evaluate the condition for the onset of anticrack (ac) and supershear (asc) propagation. The critical crack length decreases with increasing slope angle ψ and is on the order of 0.1Λ (Fig. 3). The super-critical crack length only exists if the slope angle ψ is larger than the effective friction angle \(\phi ^* \). It also decreases with increasing slope angle and varies between 0.5Λ and 15Λ in the simulations. The effective friction coefficient controls the onset of the supershear transition and significantly depends on the collapse amplitude h of the weak layer (Fig. 3, inset). Without volumetric collapse, the effective friction angle is exactly equal to the friction angle. However, increasing collapse heights reduce the effective frictional resistance of the shear band, as reported in ref. 4. This local friction reduction enables a supershear transition for slope angles lower than the weak-layer friction angle. In effect, once the crack reaches its super-critical crack length, its sharp acceleration is associated with a significant increase of the slab section that is not supported by the weak layer, leading to unstable propagation even below the friction angle. The simulation data are well reproduced by equation (1) for an effective friction angle between 0.4ϕ and ϕ.
The asymptotic crack propagation speeds obtained in all simulations are shown in Fig. 4a. For slope angles lower than the effective friction coefficient, the asymptotic propagation speed is sub-Rayleigh and varies between 0.25cs and 0.6cs. In this regime, the peak shear-to-normal stress ratio, μp, is low due to large normal stresses at the peak (Fig. 1c) and changes in sign according to the shear stress nature (that is, negative when slab bending is dominant and positive when slab tension is dominant). For larger slope angles, the propagation becomes supershear with a speed approaching a value of \(1.6c_\rms \sim \sqrtE/\rho \), similar to the longitudinal elastic wave speed in a beam. In this case, the large values of μp indicate that this regime is driven by large shear stress and low normal stress values (Figs. 1d and 2) resulting from slab tension.
a, Asymptotic crack speeds obtained in numerical simulations. Points are coloured according to the peak shear-to-normal stress ratio μp. The background colour represents different propagation regimes limited by the Rayleigh wave speed cR, the shear wave speed cs, the Eshelby wave speed \(c_\rmE=\sqrt2c_\rms\) and the p-wave speed cp. b, Crack speeds obtained in field experiments (flat PST, dark blue triangles; 37 PST, red triangles) and full-scale avalanche measurements (cross-slope speed, blue squares; down-slope speed, red circles). The determination of error bars is described in the Methods. The inset represents the normalized slope angle versus the azimuthal angle θ, characterizing crack direction (down-slope or cross-slope). c, Location of the measurement points on the slope prior to the avalanche.
Furthermore, our findings indicate that the crack propagation speed measured in small-scale experiments is not necessarily representative of crack speeds on real avalanche terrain in the down/up-slope direction (mode II). In fact, despite the different propagation mechanisms, the experimentally measured values are in good agreement with the avalanche cross-slope propagation speeds (Fig. 4b, mode III), which are theoretically limited by the Rayleigh wave speed23. Here we provide a two-dimensional (2D) theoretical framework that allows the conditions for the onset of this transition to be evaluated, as well as the crack propagation speed, which depends on slab elastic waves. Future work should include slope-scale experiments and simulations to study 3D propagation patterns35,36, as well as the complex interplay between the weak layer and slab fracture during the release process.
Our findings shed light on a previously unreported stage of the avalanche release process, with key implications for predictions of avalanche danger. More generally, our results reinforce the analogy between snow slab avalanches and earthquakes. Although the mechanism of supershear propagation has rarely been reported in large strike-slip earthquakes15, it requires a very common combination of topographical and mechanical ingredients in slab avalanche release.
On 31 January 2019, a professional snowboarder triggered a dry-snow slab avalanche (Extended Data Fig. 4a) in a location near Col du Cou in Wallis, Switzerland (Extended Data Fig. 5a). A few minutes previously, the group had checked the snowpack stability on a slope immediately behind and did not trigger an avalanche (see ski tracks in Extended Data Fig. 4b). The snowboarder triggered this slab avalanche because of the large impact force induced by a jump from the ridge (Supplementary Video 3), which led to failure initiation and crack propagation in the buried weak snow layer close to the ground.
The avalanche was recorded using a high-quality video camera (with a frame rate of 50 frames per second), allowing analysis of slab motion induced by crack propagation within the buried weak snow layer. Our crack propagation speed measurements rely on four stages: (1) video stabilization using optical flow, (2) Eulerian video magnification (EVM) to enhance small changes in snow reflection due to slab deformation, and detection of (3) time and (4) location of slab deformation between video frames. After determining the location and timing of slab deformation, we calculated the deformation distances and time difference from the crack initiation point (impact of the snowboarder) and time and accounted for spatial and temporal uncertainties. This allowed us to drive an estimation of the average crack propagation speed in the determined direction.
Our procedure to evaluate the crack propagation speed based on the EVM method was validated by applying it to (1) 2D and 3D numerical simulations performed on flat and inclined slopes (Extended Data Fig. 6) and (2) classical PST experiments on both flat and inclined slopes and a flat 5-m-long PST31 (Extended Data Fig. 7).
After the supershear transition, the crack propagates in mode II (Fig. 1). We thus describe the onset of supershear transition based on a shear band propagation model to predict the supercritical crack length. The approach used is similar to that in refs. 29,47,48, but introduces the effect of the collapse height on the residual shear friction.
We thank R. Flück for the avalanche photographs. We also thank M. Schaer, L. De Martin and T. Bessire for providing us with additional information and materials regarding the slab avalanche in Col de Cou. We acknowledge S. Mayer for the evaluation of the slab density in the full-scale avalanche based on manual snow profiles and SNOWPACK simulations. We acknowledge J.-F. Molinari for helpful and constructive discussions on the topic of supershear crack propagation. J.G. acknowledges financial support from the Swiss National Science Foundation (SNF; grant no. PCEFP2_181227). G.B. and B.B. were supported by a grant from the Swiss National Science Foundation (200021_169424). C.J. acknowledges financial support from the National Science Foundation (awards nos. 2153851, 2153863 and 2023780) and the Department of Energy (award no. ORNL 4000171342) of the United States. 2ff7e9595c
Comments